Tim the Enchanter

There are some who call me.... Tim.

It will occur to anyone who plays enough Resistance or Avalon that there's a great deal of probability afoot. If you've ever said "There's a 50/50 chance this team is bad!" you know what I'm talking about.

Tim the Enchanter is a tool to track these games, build a (very simplistic) Bayesian model of the situation as it stands, and, hopefully, deduce exactly what everyone's role is. In the future, it may also turn into a full-blown AI.

Written in Python (fortunately thematic!).

So how did you become king, then?

git clone https://github.com/barakmich/tim-the-enchanter.git
cd tim-the-enchanter
pip install -r requirements.txt
python tim.py

That'll get you going. I highly recommend you also create a virtualenv using virtualenvwrapper so that you don't muck with your system install of Python.

You may notice it's slow on 10-player games. For this, I recommend using pypy; tim-the-enchanter is compatible with pypy v1.9 and higher, which is convienent for Ubuntu users, who can simply

apt-get install pypy
pypy tim.py

Or likewise, choose pypy for their virtualenv.

If she weighs the same as a duck, she's made of wood. And therefore..... a witch!

So you're running tim.py. Now what?

help

Can give you a list of commands. Most commands ask followup questions. Until I get a better parse going, you'll have to work in binary and give the players at your table numbers, 0-9. Here's an example exchange:

Tim the Enchanter v1.0
None> newgame
How many players? 5
5 Player Game (0 constraints)> vote
Team? 12
Votes? 10000
Round? 1
# Fails Required? 1
5 Player Game (1 constraints)> eval 300000
Simulating games: |*********************************************************| Time: 0:00:08
 (4): 60.861300% Good 39.138700% Evil
 (3): 60.839685% Good 39.160315% Evil
 (1): 59.810877% Good 40.189123% Evil
 (2): 59.801591% Good 40.198409% Evil
 (0): 58.686548% Good 41.313452% Evil
5 Player Game (1 constraints)> 

We started a new game, and player zero voted for a team that everyone else hated and he wasn't on. That makes him look a little evil (and casts a little suspicion on 1 and 2, who were on the team). But not too much -- it's only round one. Suppose he had done that on round 5 instead?

5 Player Game (1 constraints)> ls    
0: Vote -- Team: [1, 2] Votes: [1, 0, 0, 0, 0] Round: 1 
5 Player Game (1 constraints)> disb 0
5 Player Game (0 constraints)> vote
Team? 12
Votes? 10000
Round? 5
# Fails Required? 1
5 Player Game (1 constraints)> eval 300000
Simulating games: |********************************************************| Time: 0:00:04
 (4): 66.937259% Good 33.062741% Evil
 (3): 66.898706% Good 33.101294% Evil
 (1): 61.507670% Good 38.492330% Evil
 (2): 61.462866% Good 38.537134% Evil
 (0): 43.193499% Good 56.806501% Evil

More clearly evil, just from that one vote. In concert with the other assertions, the possibilities will help bring the truly evil to light.

"You know much that is hidden, oh Tim." "Quite."

Feel free to help me improve it. Do drop me a line, or follow me on Twitter. Special thanks to Nyeek game nights where I've been known to field test this! If you're in NYC, come visit!